A Comparative Study on Hashing Algorithms for Data Integrity and Efficiency

In recent years, the widespread availability of image editing tools has led to the proliferation of phony and manipulated photos on the Internet and social media. Various techniques have been developed to detect image forgery and identify altered or fabricated regions, with a growing emphasis on deep learning (DL) methods. This study explores recent advances in DL-based forgery detection algorithms, focusing on the detection of copy-move and splicing attacks two of the most common image tampering techniques. Additionally, the challenges posed by DeepFake-generated content, which often mimics splicing manipulation, are discussed. The study also compares hashing algorithms (SHA-256, CRC32, Random Projection Hashing, and Count-Min Sketch) for use in data integrity, similarity searches, and frequency estimation. Finally, recommendations for selecting suitable algorithms and hybrid approaches are provided to enhance image authentication and large-scale data analysis.